Rob Edwards' article Safety warning at nuclear bomb plant in the 2007-09-22 issue of NewScientist (p. 8-9) shows a diagram of a bunker-like building used to dismantle nuclear warheads. The figure shows the high-explosives in the warhead exploding, pushing up the ceiling and weakening it. Then, it shows the ceiling coming down and collapsing, with the gravel originally atop the ceiling coming down and burying the remains. This design is supposed to mitigate dispersal of dust containing nuclear material.
But, it makes one wonder: If the volume of gasses in the room is greatly increased due to the detonation of the high explosive components of the warhead, where do those gases go when the roof with all its gravel caves in on the remains of the warhead? Wouldn't they escape through the ruptured ceiling and gravel during the collapse? And, wouldn't that gas release through the gravel covering during collapse be sufficiently energetic to carry with it the very material the whole apparatus was supposed to prevent escaping?
And so, shouldn't we be having these facilities rated for a particular rate and amount of expansion in case there is a detonation, and design them to have an attached low-pressure area of sufficient volume that the resulting volume of gasses from the explosion would be contained with much more assurance by expanding into that space? As with the ceiling in the design shown, the partition between the dismantling room and the low-pressure area could be set up to be strong enough to maintain the pressure differential under normal circumstances, yet weak enough that an explosion during work on a warhead would rupture the barrier and allow the explosion products to be at least partially contained by expansion into the previously low-pressure chamber.
That's My 1/50th.